Abstract
Recent molecular studies indicate that aerobic glycolysis plays an important role in tumorigenesis and is a valid target for cancer therapy. Although 2-deoxyglucose (2-DG) is well characterized as a glycolytic inhibitor, we recently discovered that it activates a prosurvival oncoprotein, AKT, through PI3K. In this study, we discovered that 2-DG treatments disrupted the binding between insulin-like growth factor 1 (IGF-1) and IGF-binding protein 3 (IGFBP3) so that the free form of IGF-1 could be released from the IGF-1·IGFBP3 complex to activate IGF-1 receptor (IGF1R) signaling. Because IGF1R signaling is involved, PI3K/AKT constitutes only one of the prosurvival pathways that are activated by 2-DG treatment; we validated that MEK-ERK signaling was also induced in an IGF1R-dependent manner in some cancer cell lines. Furthermore, our phospho-specific antibody microarray analysis indicated that 2-DG up-regulated the phosphorylation of 64 sites within various signaling pathways in H460 cells. Chemical inhibition of IGF1R reduced 57 of these up-regulations. These data suggest that 2-DG-induced activation of many survival pathways can be jointly attenuated through IGF1R inhibition. Our in vitro analysis demonstrated that treatment with a combination of subtoxic doses of 2-DG and the IGF1R inhibitor II reduced cancer cell proliferation 90% and promoted significant apoptosis.
Cancer cells display high rates of aerobic glycolysis in comparison with their nontransformed counterparts (i.e. the Warburg effect (1)). Whether increased aerobic glycolysis drives tumor formation or merely represents a byproduct of oncogenic transformation has been a subject of controversy. Two recent studies demonstrated that the Warburg effect can be reversed in some cancer cells by either the depletion of lactate dehydrogenenase A or switching pyruvate kinase expression from M2 to M1 isoform (2, 3). Interestingly, the reversal of the Warburg effect correlates with a reduction in the ability of the isogenic cancer cells to form tumors in nude mouse xenografts. Viewed in combination, these observations appeared to indicate that tumor cells preferentially use glucose for purposes other than oxidative phosphorylation and that aerobic glycolysis is a valid target for cancer therapeutics.
Targeting glycolysis for cancer treatment has been explored previously as a therapeutic approach (4, 5). Of all the glycolysis inhibitors that were evaluated, 2-deoxyglucose (2-DG)3 is the one that has been best characterized in animal model studies and human clinical trials (6–8). It is converted by hexokinase to phosphorylated 2-DG, which becomes trapped inside the cell and inhibits hexokinase (9). As a direct consequence of 2-DG treatment, intracellular ATP is depleted (10, 11), which ultimately suppresses cell proliferation in vitro (12, 13). Nonetheless, the implementation of 2-DG as an anticancer agent in vivo has been a disappointment. Whereas 2-DG suppresses cell growth in vitro, studies using xenografts indicate that 2-DG treatment, when provided as a single agent, does not inhibit tumor growth (6).
Because 2-DG is a small molecule, we suspected that it activates other signaling pathways and decided to evaluate its off-target effects. Our initial findings indicated that 2-DG activates AKT function through phosphatidylinositol 3-kinase (PI3K) and is independent of glycolysis or mTOR inhibition. Thus, the inhibitory effect on growth produced by 2-DG-mediated glycolysis inhibition may be partial offset by the fact there is also 2-DG-induced AKT activation (14). In the current study, we used a phospho-specific antibody array to identify IGF1R as the upstream receptor tyrosine kinase that is responsible for the activation of AKT signaling. Using recombinant IGF-1 and IGFBP3 proteins, we discovered that the inhibition of IGF-1 by IGFBP3 is disrupted in the presence of 2-DG. As 2-DG treatment activates IGF-1 signaling, we evaluated other prosurvival signaling pathways such as ERK signaling, which was also activated by 2-DG treatment in some cancer cell lines. Lastly, we tested to see whether an inhibitor of IGF1R would interfere with the prosurvival pathways and increase apoptosis if given in combination with 2-DG.
EXPERIMENTAL PROCEDURES
Materials
2-Deoxyglucose, rabbit polyclonal anti-actin antibody, and recombinant IGF-1 were purchased from Sigma-Aldrich. Erlotinib, LY294002, U0126, and PB98059 (MEK inhibitors) were purchased from LC Laboratories (Woburn, MA). IGF1R inhibitor II (catalog No. 407248), αIR3 (an IGF1R antibody), and PP2 were purchased from Calbiochem. Recombinant IGFBP3 was purchased from GenWay Biotech Inc. (San Diego, CA). An active free IGF-1 ELISA kit was purchased from Beckman Coulter. Antibodies against AKT, phospho-AKT (p-AKT-Ser473), ERK, phospho-ERK (p-p44/42/Thr202/Tyr204), caspase-3, and IGF1R were purchased from Cell Signaling Technology, Inc. (Beverly, MA).
Cell Lines and Cell Culture
Non-small cell lung cancer (NSCLC) cell lines (H460, A549, H157, H23, H1299, H1792, H358, Calu-1, H226, H522, HOP92, H596, and H322M), breast cancer cell line (T-47D), cervical cancer cell line HeLa, melanoma cell line MDA-MB-435, and colorectal cancer cell line HCT116 were purchased from the American Type Culture Collection (ATCC, Manassas, VA) and were propagated according to the conditions recommended by ATCC.
Phospho-specific Protein Microarray Analysis
A phospho-specific protein microarray was obtained from Full Moon Biosystems, Inc. Protein microarray analysis was carried out using a previously described protocol, which is included in the supplemental “Methods” (15).
Western Blot Analysis
The procedures for preparation of whole cell protein lysates and Western blotting were performed as described previously (14). The same blots were used for probing phospho-specific antibodies and antibodies against total protein. Actin was used as the loading control. The Western analyses present in this study were carried out at least twice, and representative images are shown in Figs. 1–5. For quantitation analysis, the density of phospho-AKT, total AKT, phospho-ERK, and total ERK was determined by using ImageJ from the National Institutes of Health (rsbweb.nih.gov/ij/). The phosphorylation ratio was calculated using the following formula.
![]() |
FIGURE 1.
2-DG-induced AKT phosphorylation is mediated through IGF1R but not EGFR or Src. A, H460 and H1299 cells were pretreated with 25 μm erlotinib for 30 min before the addition of 25 mm 2-DG. B, H460 and H157 cells were pretreated with 10 μm PP2 for 30 min before the addition of 2-DG. C, inhibition of IGF1R by a small molecule, IGF1R inhibitor II. H1299, H157, H460, A549, and Calu-1 cells were pretreated with 10 μm IGF1R inhibitor II for 30 min before the addition of 2-DG. D, inhibition of IGF1R by an IGF1R monoclonal antibody (αIR3). H460 and H1299 cells were pretreated with 5 μg/ml αIR3 for 30 min before the addition of 2-DG. E, siRNA against IGF1R was used to deplete the expression of IGF1R in H460 cells, and anti-IGF1R antibody was used to evaluate the extent of IGF1R depletion. Cells were harvested 4 h after 25 mm 2-DG treatment and analyzed by immunoblot. Band intensity was quantified by densitometry, and phosphorylation changes were expressed by a ratio using the formula described under “Experimental Procedures.”
FIGURE 2.
2-DG alleviates the suppression of IGF-1-induced AKT phosphorylation by IGFBP3. A, H460 were treated with 25 mm 2-DG in the presence of 10, 1, or 0% fetal bovine serum (FBS). B, H460 cells were serum-starved for 12 h before being treated with 25 mm 2-DG, 20 ng/ml recombinant IGF-1, 200 ng/ml recombinant IGFBP3, or indicated combinations. Cells were harvested 4 h after 25 mm 2-DG treatment and analyzed by immunoblot. Anti-phospho-AKT analysis was shown in both a short exposure (SE) and long exposure (LE). Phosphorylation changes were expressed as ratios based on densitometry analysis of long exposure blots. C, equal amounts of free IGF-1 (100 pg) were incubated alone, with 20 ng of IGFBP3 or 25 mm 2-DG, or with 20 ng of IGFBP3 and the indicated concentration of 2-DG. The free form of IGF-1 was evaluated using a free IGF-1 ELISA kit. Reactions were carried out in duplicate, and error bars represent 1 S.D.
FIGURE 3.
Induction of ERK phosphorylation by 2-DG treatment. A, ERK phosphorylation in H1299, H460, and H1157 cell lines after 15 min, 30 min, 1 h, and 2 h of treatment with 25 mm 2-DG. B, ERK phosphorylation after treatment with 25 mm 2-DG in H1299, H460, and H1157 cell lines as measured at 8 and 24 h. C, ERK phosphorylation in H157, H1299, and H460 cell lines after a 2-h treatment with 2.5 mm 2-DG. Phosphorylation changes were expressed as ratios based on densitometry analysis. D, ERK phosphorylation in A549, H23, MDA-MB-435, H1792, and H358 cells after a 2-h treatment with 25 mm 2-DG.
FIGURE 4.
2-DG induces ERK phosphorylation through IGF1R and MEK. A, H1299, H460, and H157 cells were pretreated with either 10 μm U0126 or 20 μm PD98059 for 30 min before the addition of 25 mm 2-DG. B, H1299, H157, H460, and MDA-MB-435 cells were pretreated with 10 μm IGF1R inhibitor II for 30 min before the addition of 2-DG. C, H460 and H157 cells were pretreated with 5 μg/ml IGF1R monoclonal antibody (αIR3) for 30 min before the addition of 2-DG. D, siRNA against IGF1R was used to deplete the expression of IGF1R in H157 cells, and anti-IGF1R antibody was used to evaluate the extent of IGF1R depletion. Cell lysates were collected 2 h after 2-DG treatment and probed with anti-phospho-p44/42 MAPK(Thr202/Tyr204) antibody using total ERK and actin as loading controls (Ctrl). Phosphorylation changes were expressed as ratios based on densitometry analysis.
FIGURE 5.
The combined inhibition of glycolysis and IGF1R induce synergistic cell killing. A, H460 and H157 cells were seeded in 96-well plates and treated with either 10 μm LY294002 alone or the indicated concentration of U0126 or IGF1R inhibitor II alone, 5 mm 2-DG alone, and 2-DG in combination with either inhibitor. Plates were subjected to methanethiosulfonate assay after 24 h (H157) or 48 h (H460). Reactions were carried out in quadruplicate, and error bars represent 1 S.D. B, H460 and H157 cells were seeded in 6-well plates and pretreated with 10 μm LY294002, 10 μm U0126, or 10 μm IGF1R inhibitor II before the addition of 5 mm 2-DG. Cell lysates were collected at 4 and 48 h after treatment and probed with anti-phospho-AKT (Ser473) and anti-phospho-p44/42 MAPK(Thr202/Tyr204) antibodies (4 h) or anti-caspase-3 antibody (48 h). Total AKT, ERK, and actin were used as loading controls. C, annexin V and 7-AAD analysis of apoptosis. H460 and H157 cells were seeded in 6-well plates and treated with 10 μm IGF1R inhibitor II (IGF-1Ri) alone, 5 mm 2-DG alone, or a combination of both. Both floating and attached cells were collected 48 h after treatment and subjected to flow analysis.
Transient siRNA Transfection
IGF1R siRNA duplexes were purchased from Applied Biosystems (Foster City, CA). Transient IGF1R knockdown was performed as described previously (also included in supplemental “Methods”) (14).
Active Free IGF-1 Analysis
100 pg of free IGF-1 was incubated with 20 ng of IGFBP3 in a 10-μl volume at 4 °C overnight to form the IGF-1·IGFBP3 complex. This complex was subjected to various concentrations of 2-DG for 15 min at room temperature, and the free form of IGF-1 was determined by the active free IGF-1 ELISA kit using a protocol provided by the manufacturer.
Cell Growth Inhibition Assay
2 × 103 H460 or H157 cells were seeded in 96-well cell culture plates. Cells were treated with 5 mm 2-DG only, 5 or 10 μm IGF1R inhibitor II only, or a combination of 2-DG and IGF1R inhibitor II. Cell growth inhibition was determined after 48 h by the CellTiter 96® AQueous nonradioactive cell proliferation assay (Promega, Madison, WI) according to the manufacturer's instruction.
Apoptosis Analyses
Apoptosis was measured using the annexin V-PE apoptosis detection kit (BD Biosciences) followed by flow cytometry. 2 × 105 H460 or H157 cells were seeded in 6-well cell culture plates. Cells were treated with 5 mm 2-DG alone, 10 μm IGF1R inhibitor II alone, or both 2-DG and IGF1R inhibitor II. Both floating and attached cells were collected 48 h after treatment, washed twice with cold phosphate-buffered saline, and suspended in 1× binding buffer. A 100-μl aliquot of the cell suspension (representing 5 × 105 cells) was transferred to a culture tube to which 5 μl of annexin V-PE and 5 μl of 7-AAD were added, and the mix was incubated for 15 min at room temperature in the dark. Apoptosis analysis was carried out using a FACScan (BD Biosciences) and FlowJo software, version 7.2. A total of 10,000 cells was collected for each sample for analysis.
RESULTS
Phospho-antibody Microarray Analysis of 2-DG-induced Protein Phosphorylation
We demonstrated previously that 2-DG treatment activates the AKT function in seven NSCLC, one colorectal cancer cell line (HCT116), one melanoma cell line (M14/MDA-MB-435), two breast cancer cell lines (T-47D and MCF-7), and one cervical cancer cell line (HeLa) (14). In the current study, we also discovered 2-DG-induced AKT phosphorylation in six more NSCLC cell lines (supplemental Fig. 1A); therefore, the induction of AKT phosphorylation by 2-DG is a common phenomenon in NSCLC and other cancer cell lines.
To identify the upstream molecules that are responsible for 2-DG-induced AKT activation, we used a customized phospho-specific antibody microarray from Full Moon Biosystems. This array interrogates phosphorylation sites on 15 receptor/cytoplasmic tyrosine kinases/tyrosine kinase adaptors, as well as 100 phosphorylation sites in various downstream signaling pathways. Each site was evaluated using a pair of antibodies against both the phosphorylated and unphosphorylated site, and every antibody was spotted in replicates of six. Based on these features, a software program (PANDA) was developed that would allow us to quantitatively evaluate the phosphorylation change at each site with a 95% confidence interval (see supplemental “Methods”). H460 cells treated with 2-DG responded with a statistically significant up-regulation of seven receptor/cytoplasmic tyrosine kinases (Table 1). In addition to confirming the activation of PI3K/AKT signaling by 2-DG, this method allowed us to determine that 2-DG treatment also induced the phosphorylation of Raf-MEK-ERK kinases or their targets (10 sites/9 proteins), cell cycle/DNA damage checkpoint proteins (6 sites/5 proteins), and JAK/STAT proteins (5 sites/5 proteins). In total, we discovered that the phosphorylation of 64 sites (56%) became elevated following 2-DG treatment (Table 1). Therefore, the off-target effects of 2-DG were broader than we had initially anticipated.
TABLE 1.
Alteration of protein phosphorylation by 2-DG, 2-DG and IGF1R inhibitor II (IGF1Ri), and IGF1R inhibitor II alone
Phosphorylation sites | 2-DG only |
2-DG + IGF1Ri |
IGF1Ri only |
Signaling pathways | |||
---|---|---|---|---|---|---|---|
Ratio | 95% CIa | Ratio | 95% CIa | Ratio | 95% CIa | ||
Caveolin-1 (phospho-Tyr14) | 2.18 | 1.64–2.71 | 0.64 | 0.54–0.74 | 1.39 | 1.04–1.75 | Cytoskeletal signaling |
Estrogen receptor-α (phospho-Ser167) | 2.11 | 1.73–2.48 | 1.35 | 1.14–1.55 | 1.82 | 1.51–2.14 | Akt, p90RSK target |
STAT1 (phospho-Tyr701) | 2.06 | 1.58–2.53 | 0.57 | 0.44–0.71 | 1.16 | 0.87–1.44 | JAK/STAT |
CREB (phospho-Ser133) | 1.83 | 1.63–2.04 | 1.34 | 1.18–1.49 | 1.39 | 1.20–1.58 | ERK target |
IGF1R (phospho-Tyr1161) | 1.80 | 1.59–2.01 | 1.20 | 1.06–1.35 | 1.36 | 1.16–1.55 | Receptor/cytoplasmic tyrosine kinases |
MKK3 (phospho-Ser189) | 1.78 | 1.32–2.23 | 0.92 | 0.68–1.15 | 0.99 | 0.58–1.40 | MAPK |
ICAM-1 (phospho-Tyr512) | 1.76 | 1.54–1.98 | 0.99 | 0.86–1.12 | 1.96 | 1.50–2.43 | Adhesion |
NF-κB-p105/p50(phospho-Ser893) | 1.72 | 1.39–2.04 | 0.75 | 0.59–0.90 | 2.24 | 1.85–2.62 | NF-κB |
JunD (phospho-Ser255) | 1.70 | 1.50–1.89 | 1.52 | 1.33–1.70 | 1.39 | 1.26–1.51 | |
Tau (phospho-Ser404) | 1.67 | 1.45–1.89 | 0.70 | 0.59–0.82 | 1.15 | 0.94–1.36 | ERK target |
Src (phospho-Tyr418) | 1.63 | 1.17–2.08 | 1.04 | 0.75–1.34 | 1.95 | 1.40–2.50 | Receptor/cytoplasmic tyrosine kinases |
BAD (phospho-Ser136) | 1.58 | 1.35–1.81 | 1.30 | 1.14–1.46 | 2.02 | 1.73–2.30 | Apoptosis |
GSK3-β (phospho-Ser9) | 1.58 | 1.34–1.82 | 0.92 | 0.79–1.05 | 3.63 | 2.45–4.81 | PI3K/AKT |
JunB (phospho-Ser79) | 1.58 | 1.40–1.75 | 0.44 | 0.36–0.51 | 1.44 | 1.17–1.72 | Lymphocyte signaling |
HER2 (phospho-Tyr877) | 1.58 | 1.41–1.75 | 1.09 | 0.98–1.20 | 1.11 | 1.00–1.22 | Receptor/cytoplasmic tyrosine kinases |
Chk1 (phospho-Ser317) | 1.55 | 1.35–1.75 | 0.71 | 0.63–0.79 | 0.95 | 0.81–1.10 | Cell cycle/checkpoint |
p53 (phospho-Ser315) | 1.53 | 1.37–1.69 | 0.95 | 0.85–1.06 | 1.31 | 1.07–1.55 | Cell cycle/checkpoint |
Myc (phospho-Ser373) | 1.48 | 1.33–1.64 | 0.43 | 0.36–0.50 | 0.94 | 0.83–1.04 | Apoptosis/autophage |
BAD (phospho-Ser112) | 1.48 | 1.16–1.80 | 0.98 | 0.64–1.32 | 1.42 | 1.04–1.79 | Apoptosis |
JAK2 (phospho-Tyr1007) | 1.48 | 1.17–1.79 | 0.88 | 0.67–1.08 | 0.90 | 0.61–1.19 | JAK/STAT |
TYK2 (phospho-Tyr1054) | 1.47 | 1.27–1.67 | 0.61 | 0.48–0.75 | 1.52 | 1.25–1.79 | Receptor/cytoplasmic tyrosine kinases |
Integrin β-3 (phospho-Tyr773) | 1.46 | 1.32–1.59 | 0.96 | 0.86–1.05 | 1.47 | 1.32–1.63 | Adhesion |
p53 (phospho-Ser6) | 1.45 | 1.26–1.64 | 0.90 | 0.80–1.01 | 1.98 | 1.75–2.22 | Cell cycle/checkpoint |
CDC2 (phospho-Tyr15) | 1.45 | 1.27–1.63 | 1.17 | 0.99–1.36 | 1.96 | 1.73–2.19 | Cell cycle/checkpoint |
P38 MAPK (phospho-Tyr182) | 1.44 | 1.19–1.70 | 1.21 | 1.04–1.37 | 2.13 | 1.76–2.50 | MAPK |
Rb (phospho-Ser780) | 1.44 | 1.30–1.58 | 0.93 | 0.86–1.00 | 0.87 | 0.81–0.94 | Cell cycle/checkpoint |
eEF2K (phospho-Ser366) | 1.42 | 1.23–1.62 | 0.96 | 0.87–1.06 | 0.63 | 0.56–0.70 | Translational control |
FKHR (phospho-Ser256) | 1.42 | 1.28–1.57 | 0.82 | 0.74–0.90 | 1.02 | 0.92–1.12 | PI3K/AKT |
TrkB (phospho-Tyr515) | 1.42 | 1.24–1.60 | 0.83 | 0.70–0.97 | 0.90 | 0.80–1.01 | Receptor/cytoplasmic tyrosine kinases |
PDK1 (phospho-Ser241 | 1.41 | 1.19–1.63 | 0.83 | 0.59–1.07 | 1.19 | 0.99–1.38 | PI3K/AKT |
I-κ-B-ϵ (phospho-Ser22) | 1.38 | 1.16–1.59 | 0.84 | 0.71–0.96 | 1.90 | 1.73–2.06 | NF-κB |
Chk2 (phospho-Ser516) | 1.37 | 1.15–1.60 | 0.59 | 0.50–0.67 | 0.65 | 0.56–0.74 | Cell cycle/checkpoint |
p44/42 MAPK (phospho-Thr202) | 1.37 | 1.23–1.51 | 1.22 | 1.09–1.35 | 1.85 | 1.63–2.07 | MAPK |
p44/42 MAPK (phospho-Tyr204) | 1.35 | 1.12–1.57 | 1.19 | 0.93–1.44 | 1.42 | 1.26–1.57 | MAPK |
BCL-XL (phospho-Ser62) | 1.34 | 1.02–1.65 | 1.12 | 0.84–1.39 | 2.04 | 1.59–2.49 | Apoptosis |
Shc (phospho-Tyr349) | 1.33 | 1.07–1.58 | 0.94 | 0.77–1.11 | 0.97 | 0.75–1.19 | Tyrosine kinase/adaptors |
EGFR (phospho-Tyr1110) | 1.33 | 1.13–1.52 | 1.11 | 0.84–1.38 | 1.05 | 0.95–1.14 | Receptor/cytoplasmic tyrosine kinases |
BRCA1 (phospho-Ser1524) | 1.33 | 1.08–1.57 | 1.75 | 1.06–2.44 | 0.43 | 0.34–0.51 | DNA damage checkpoint |
VEGFR2 (phospho-Tyr951) | 1.32 | 1.07–1.56 | 0.94 | 0.77–1.12 | 0.93 | 0.75–1.11 | Receptor/cytoplasmic tyrosine kinases |
JAK2 (phospho-Tyr221) | 1.32 | 1.13–1.51 | 0.76 | 0.65–0.87 | 0.96 | 0.84–1.08 | JAK/STAT |
CaMKII (phospho-Thr286) | 1.31 | 1.09–1.52 | 0.75 | 0.65–0.85 | 0.77 | 0.70–0.84 | Stress |
c-Jun (phospho-Ser243) | 1.29 | 1.11–1.48 | 1.27 | 1.07–1.47 | 1.24 | 1.09–1.39 | MAPK |
eIF2-α (phospho-Ser51) | 1.29 | 1.10–1.49 | 0.98 | 0.69–1.27 | 1.08 | 0.92–1.24 | Translational control |
GSK3-α (phospho-Ser21) | 1.29 | 1.17–1.40 | 0.75 | 0.64–0.85 | 2.04 | 1.75–2.32 | PI3K/AKT |
SHP-2 (phospho-Tyr580) | 1.29 | 1.11–1.46 | 0.99 | 0.82–1.15 | 0.89 | 0.77–1.01 | MAPK |
AMPK1 (phospho-Thr174) | 1.28 | 1.04–1.52 | 1.00 | 0.80–1.19 | 1.40 | 1.12–1.68 | Stress |
c-Kit (phospho-Tyr721) | 1.28 | 1.19–1.37 | 1.58 | 1.31–1.85 | 1.75 | 1.63–1.88 | Receptor/cytoplasmic tyrosine kinases |
p70 S6 kinase (phospho-Ser424) | 1.28 | 1.09–1.46 | 0.92 | 0.75–1.09 | 0.89 | 0.79–0.99 | Translational control |
Integrin β-3 (phospho-Tyr785) | 1.27 | 1.17–1.37 | 1.03 | 0.95–1.10 | 1.86 | 1.59–2.13 | Adhesion |
JunB (phospho-Ser259) | 1.27 | 1.13–1.40 | 1.16 | 1.05–1.28 | 1.17 | 1.07–1.28 | Lymphocyte signaling |
MEK1 (phospho-Thr291) | 1.26 | 1.08–1.44 | 1.08 | 0.86–1.30 | 1.20 | 1.06–1.34 | MAPK |
STAT5A (phospho-Tyr694) | 1.23 | 1.02–1.44 | 1.16 | 0.91–1.40 | 0.54 | 0.47–0.62 | JAK/STAT |
Chk1 (phospho-Ser280) | 1.21 | 1.03–1.40 | 0.99 | 0.83–1.15 | 1.43 | 1.20–1.67 | Cell cycle/checkpoint |
MEK-2 (phospho-Thr394) | 1.21 | 1.11–1.31 | 0.76 | 0.69–0.84 | 1.60 | 1.49–1.70 | MAPK |
4E-BP1 (phospho-Thr36) | 1.21 | 1.04–1.38 | 1.31 | 0.71–1.90 | 1.19 | 1.05–1.32 | Translational control |
SAPK/JNK (phospho-Thr183) | 1.20 | 1.12–1.28 | 1.10 | 0.86–1.35 | 1.05 | 0.96–1.13 | MAPK |
Myc (phospho-Thr58) | 1.20 | 1.07–1.32 | 0.56 | 0.44–0.68 | 1.18 | 0.92–1.45 | Apoptosis/autophage |
Pyk2 (phospho-Tyr402) | 1.19 | 1.06–1.33 | 1.33 | 0.80–1.86 | 1.13 | 1.00–1.26 | Receptor/cytoplasmic tyrosine kinases |
NF-κB-p65 (phospho-Ser529) | 1.18 | 1.02–1.33 | 1.33 | 0.68–1.98 | 1.10 | 0.97–1.23 | NF-κB |
14-3-3 ζ (phospho-Ser58) | 1.18 | 1.07–1.28 | 0.98 | 0.86–1.10 | 1.26 | 1.06–1.46 | Tyrosine kinase/adaptors |
Raf1 (phospho-Ser259) | 1.18 | 1.08–1.27 | 1.07 | 0.96–1.19 | 0.91 | 0.79–1.03 | MAPK |
Rac1/cdc42 (phospho-Ser71) | 1.17 | 1.02–1.33 | 1.65 | 0.86–2.44 | 1.20 | 1.06–1.34 | Cytoskeletal signaling |
Elk-1 (phospho-Ser383) | 1.15 | 1.05–1.24 | 1.16 | 1.10–1.23 | 1.20 | 1.09–1.30 | MAPK |
JAK1 (phospho-Tyr1022) | 1.14 | 1.03–1.25 | 0.77 | 0.67–0.86 | 0.86 | 0.77–0.95 | JAK/STAT |
a CI, confidence interval.
2-DG Induces AKT Phosphorylation through IGF1R but Not EGFR or Src
Because 2-DG treatment enhanced the phosphorylation of several receptor tyrosine kinases, we next sought to determine whether a specific receptor is responsible for the induction of AKT phosphorylation. We first used erlotinib, an EGFR inhibitor. Because this compound failed to block 2-DG-induced AKT phosphorylation in H460 and H1299 cells (Fig. 1A), we were able to confirm in new cell lines our previous observation that in H1650 cells, EGFR was unlikely to be involved in 2-DG-induced AKT activation (14). As Src is a tyrosine kinase that mediates multiple signal transduction pathways (16), we next evaluated the effect of PP2, a Src inhibitor, on 2-DG-induced AKT phosphorylation. However, pretreatment of H460 and H157 cells with PP2 did not block 2-DG-induced AKT phosphorylation either (Fig. 1B), indicating that Src was also not required for this induction process.
Finally, we evaluated whether the inhibition of IGF1R function could inhibit 2-DG-induced AKT phosphorylation. Initially, we pretreated the NSCLC cell lines H1299, H157, H460, A549, and Calu-1 with a specific inhibitor of IGF1R, IGF1R inhibitor II, and found that this inhibitor blocked 2-DG-induced AKT phosphorylation (Fig. 1C). A similar inhibition was also observed in cell lines representing other cancer types, such as HCT116, HeLa, T-47D, and MDA-MB-435 cells (supplemental Fig. 2). To directly evaluate the role of IGF1R, we pretreated H460 and H1299 cells with an IGF1R antibody, αIR3. We found this antibody also blocked 2-DG-induced AKT phosphorylation (Fig. 1D). Alternatively, IGF1R was transiently depleted in H460 cells using siRNA, and this depletion of IGF1R expression also inhibited 2-DG-induced AKT phosphorylation (Fig. 1E). Therefore, we found that IGF1R is required for 2-DG-induced AKT phosphorylation in several cancer cell types including NSCLC.
2-DG Treatment Releases the Suppression of IGF-1-mediated AKT Phosphorylation by IGFBP3
We next sought to determine whether ligand binding is required for 2-DG-induced AKT phosphorylation. First, H460 cells were treated with 25 mm 2-DG in the presence of 10, 1, or 0% fetal bovine serum for 1 h. The level of AKT phosphorylation significantly correlated with the percentage of serum included in culture media (Fig. 2A), indicating that IGF1R ligands are required for 2-DG-induced AKT phosphorylation. Under serum-free conditions, 2-DG-induced AKT phosphorylation was observed only with extended film exposure (Fig. 2A, compare lane 5 with 6). These data indicate that H460 cells are unlikely to secrete significant amount of IGF1R ligands in response to 2-DG treatment.
IGF-1 is a major IGF1R ligand, and the binding of IGF-1 to IGF1R also activates AKT phosphorylation (17). Less than 1% of IGF-1 is present in serum as the free form because the majority of circulating IGF-1 is bound to plasma IGF-binding proteins (IGFBPs), particularly IGFBP-3 (18). The complex between IGF-1 and IGFBP-3 prolongs the half-life of IGF-1 and modulates the interaction between IGF-1 and IGF1R. We next tested how IGF-1-induced AKT was affected by 2-DG and IGFBP3 treatment. H460 cells were serum-starved for 12 h. Treatment with 200 ng/ml recombinant IGFBP3 did not alter AKT phosphorylation (Fig. 2B, lane 3). Treatment with 2-DG alone, or a mixture of 2-DG and IGFBP3 induced a similarly low level of AKT phosphorylation (Fig. 2B, lanes 2 and 7), indicating that IGFBP3 cannot suppress 2-DG-induced AKT phosphorylation. Treatment of serum-starved H460 cells with 20 ng/ml recombinant IGF-1 induced significant AKT phosphorylation (Fig. 2B, lane 4). The same amount of IGF-1 was also incubated with 250 ng/ml recombination IGFBP-3 for 12 h before the IGF-1/IGFBP-3 mixture was added to H460 cells. Pretreatment of IGF-1 with IGFBP3 significantly attenuated AKT phosphorylation (Fig. 2B, lane 5), indicating that IGFBP3 suppressed IGF-1-induced AKT phosphorylation. The addition of 2-DG to the IGF-1/IGFBP3 mixture restored AKT phosphorylation (Fig. 2B, lane 6), indicating that 2-DG treatment may disrupt the interaction between IGF-1 and IGFBP3. The promoter of the IGFBP3 gene is not methylated in H460 cells, and H460 cells secrete relatively high level of IGFBP3 into the media (19). H460-secreted IGFBP3 appeared to be able to sequester H460-screted IGF-1 or a fraction of recombinant IGF-1 because the addition of 2-DG enhanced AKT phosphorylation (Fig. 2B, compare lanes 1 and 4 with lanes 2 and 8). These data support the notion that 2-DG alleviates the suppression of IGF-1-induced AKT phosphorylation by IGFBP3.
To directly assess whether 2-DG interferes with the interaction between IGF-1 and IGFBP3, we used a free IGF-1 ELISA that was designed specifically to detect the free form IGF-1 but not the ones in the IGF-1·IGFBP3 complex. The inclusion of 25 mm 2-DG with 100 pg of IGF-1 did not affect the detection of free IGF-1 by the ELISA. 100 pg of free IGF-1 was incubated with 20 ng of recombinant IGFBP3, and the recombinant IGFBP3 sequestered the majority of IGF-1 so that only 2.2 pg of free IGF-1 could be detected in the presence of IGFBP3 (Fig. 2C). The addition of 1 mm 2-DG did not affect the formation of this complex, but the inclusion of 5 mm 2-DG increased the detectable free IGF-1 to 5.3 ng, which represent a 141% increase. Treatment with 25 mm 2-DG led to the detection of 62 pg of free IGF-1 (Fig. 2C). These data indicate that binding between IGF-1 and IGFBP3 can be disrupted by 2-DG treatment in a dose-dependent manner.
2-DG Induces Time- and Dose-dependent ERK Phosphorylation
Because the activation of IGF1R is known to induce other signaling pathways (20), and our protein array analysis revealed the up-regulation of Raf1(Ser259) MEK1(Ser221), MEK2(Thr394), and ERK(Thr202/Tyr204) phosphorylation after 2-DG treatment, we next evaluated whether 2-DG induces ERK phosphorylation. A time course analysis indicated that 2-DG treatment induced ERK phosphorylation in 15 min in the H1299 and H157 cells and in 30 min in the H460 cells (Fig. 3A). We also found that 2-DG did not transiently induce ERK phosphorylation, because it remained elevated: ERK phosphorylation continued to be observed even 8–24 h after the addition of 2-DG (Fig. 3B). It is of note that the oral administration of 300 mg/kg 2-DG will produce about 5.5 mm 2-DG in patient blood (8); our dose analysis indicated that a treatment of only 2.5 mm 2-DG was sufficient to induce ERK phosphorylation in H157, H1299, and H460 cells (Fig. 3C). Therefore, we found that 2-DG induces ERK phosphorylation in these cancer cells at a pharmacologically relevant concentration. In addition, we also observed 2-DG-induced ERK phosphorylation in A549, H23, M14/MDA-MB-435, H1792, and H358 cells (Fig. 3D). Thus, this activity by 2-DG occurs after treatment of some cancer cell lines.
2-DG Also Induces ERK Phosphorylation through IGF1R and MEK
We wished to determine whether the observed 2-DG-induced ERK phosphorylation is directly mediated by MEK. First, we evaluated the effect of U0126 in H1299 and H460 cells. The use of U0126, a highly selective inhibitor of both MEK1 and MEK2, was capable of blocking 2-DG-induced ERK phosphorylation in both cell lines (Fig. 4A, left panel). In addition, we found that 2-DG-induced ERK phosphorylation was also blocked by another MEK inhibitor, PD98059, in both H460 and H157 cells (Fig. 4A, right panel), indicating that MEK was involved in the process.
We also sought to define the role of IGF1R in 2-DG-induced ERK phosphorylation. Use of IGF1R inhibitor II also blocked ERK phosphorylation in H1299, H157, H460, and MDA-MB-435 cells following 2-DG treatment (Fig. 4B). This suggested that the increase in ERK phosphorylation was mediated by IGF1R. Furthermore, we found that IGF1R antibody blocked 2-DG-induced ERK phosphorylation in H460 and H157 cells (Fig. 4C), and the depletion of IGF1R expression in H157 cells by siRNA correlated with an attenuation of 2-DG-induced phosphorylation (Fig. 4D). Erlotinib, however, did not block 2-DG-induced ERK phosphorylation, indicating that EGFR signaling was not essential for 2-DG-induced ERK phosphorylation (data not shown). In combination, these data suggest that 2-DG-induced ERK phosphorylation is mediated through its activity on IGF1R and MEK.
Majority of 2-DG-induced Phosphorylation Is Inhibited by IGF1R Inhibitor II
We subsequently evaluated the effect of IGF1R inhibitor II on 2-DG-induced phosphorylation in the other signaling pathways found to be active with the phospho-specific antibody array. Whereas 2-DG induced the phosphorylation of IGF1R at Tyr1161 by 1.8-fold, pretreatment with IGF1R inhibitor II reduced that phosphorylation ratio down to 1.2-fold. In fact, of the 64 phosphorylation sites that were elevated by treatment with 2-DG alone, pretreatment with IGF1R inhibitor II was found to reduce phosphorylation in 57 (89%) of these sites (Table 1). These data suggest that the counterproductive off-target effects of 2-DG therapy due to enhanced protein phosphorylation could be attenuated by pretreatment with IGF1R inhibitor II. Interestingly, we also found that the basal phosphorylation levels of these 64 sites were mostly unaffected (87.5% (56 of 64)) in cancer cells that were treated solely with this inhibitor (IGF1R inhibitor II alone) (Table 1).
Combined Inhibition of Glycolysis and IGF1R Induces Synergistic Cell Killing
The inhibition of glycolysis negatively impacts cell growth (5). We had previously demonstrated that 2-DG-mediated growth suppression is partially offset by concomitant 2-DG-induced AKT activation in H1299 cells (14). Similar partial growth suppression was observed in H460 and H157 cells when 2-DG was combined with LY294002 (Fig. 5A, top panel). However, LY294002 did not alter 2-DG-induced ERK phosphorylation, so PI3K/AKT signaling was not involved in the regulation of 2-DG-induced ERK phosphorylation (Fig. 5B, top panel).
We assessed whether 2-DG-mediated growth suppression was negatively impacted specifically by 2-DG-induced ERK phosphorylation. Subtoxic doses of U0126 and 2-DG were used in this experiment. Although the combination of 5 mm 2-DG and the various doses of U0126 led to increases in growth suppression in two NSLCLC cell lines, such combinations did not substantially suppress cell proliferation (Fig. 5A, middle panel). As U0126 is specific for MEK, and we did not observe any inhibition of 2-DG-induced AKT phosphorylation, known to be mediated by PI3K (Fig. 5B), these data indicate that the activation of both AKT and ERK signaling by 2-DG did promote cell survival, but the targeted inhibition of either PI3K/AKT or MEK/ERK alone was not sufficient to substantially negate the cell survival effect induced by 2-DG.
Because IGF1R inhibitor II reversed the majority of phosphorylation sites induced by 2-DG treatment, we evaluated whether the combined inhibition of both glycolysis and IGF1R functions that were involved in promoting tumor growth would result in a synergistic inhibition of cell proliferation. We first determined the effects of combined inhibition of glycolysis and IGF1R on cell proliferation. Whereas a single dose of 2-DG (5 mm) or IGF1R inhibitor II (5 or 10 μm) resulted in ∼20–30% cell growth inhibition for H460 cells, the combination of 5 mm 2-DG and 10 μm IGF1R inhibitor II resulted in over 90% growth inhibition for these cancer cells (Fig. 5A, bottom panel). A similar synergistic effect with respect to enhanced growth inhibition was observed in H157 cells.
We also sought to determine whether apoptosis would be induced by the combination of 2-DG and IGF1R inhibitor II, because apoptosis was not observed when 2-DG was combined with LY294002 (data not shown). First, an apoptosis-based immunoblot assay was used to determine whether caspase-3 cleavage was induced by this combination. Neither the control 5 mm 2-DG nor 10 μm IGF1R inhibitor II treatments alone led to caspase-3 cleavage even 48 h after treatment, but significant caspase-3 cleavage was detected when these two compounds were used in combination (Fig. 5B, bottom panel). Additionally, apoptosis was evaluated by 7-AAD and annexin V analysis, where the combination of subtoxic doses of 2-DG and IGF1R inhibitor II resulted in a synergistic increase in the number of cells that underwent apoptosis (Fig. 5C). Therefore, the combination of subtoxic doses of 2-DG and IGF1R inhibitor not only inhibited cell proliferation but also was found to promote apoptosis.
DISCUSSION
Recent molecular biology studies indicate that aberrantly active aerobic glycolysis is required for tumorigenesis (2, 3); therefore, aerobic glycolysis is a valid target for cancer therapy. 2-DG is the best characterized glycolytic inhibitor, and the proper use of 2-DG in human subjects was previously determined in a phase I/II clinical trial in India (7, 8). As the therapeutic potential of 2-DG in combination therapy prompted enough interest in the United States, three phase I clinical trials for this compound are ongoing (ClinicalTrials.gov identifiers: NCT00096707, NCT00247403, and NCT00247403).
Because there had been limited therapeutic efficacy for 2-DG as a single agent in preclinical models, we were led to question and explore the prosurvival effects of 2-DG treatment. We initially discovered that 2-DG induces AKT activation within 15 min and that this activation is mediated by PI3K (14). Because PIK3CA and PTEN mutations are relatively rare in lung cancer (21), most NSCLC cell lines have relatively low constitutive AKT phosphorylation. Indeed, when surveyed in our current expanded analysis, we found that 87% (13 of 15) of the NSCLC cells had increases in AKT phosphorylation following 2-DG treatment.
In this study we employed a novel phospho-specific antibody microarray to identify the upstream signaling molecule/s involved in 2-DG-induced AKT activation. This antibody microarray evaluates 115 distinct phosphorylation sites in various signal transduction pathways, and the use of six replicates/antibody allowed us to identify statistically significant alterations. Our analysis revealed seven receptor tyrosine kinases as potential candidates due to significant changes in their phosphorylation, and IGF1R was subsequently verified as the upstream receptor tyrosine kinase responsible for the 2-DG-induced AKT phosphorylation that we had originally observed. Therefore, we discovered that IGF1R/PI3K/AKT constitutes a signaling pathway that becomes activated by 2-DG treatment in the majority of NSCLC cells.
Another interesting finding of this work is the mechanism by which 2-DG activates IGF1R/PI3K/AKT signaling. The majority of circulating IGF-1 is bound to plasma IGFBPs, which prolongs the half-life of IGF-1 and alters the binding of IGF-1 to IGF1R (17). Our data indicate that 2-DG treatment releases the suppression of IGF-1-mediated AKT phosphorylation by IGFBP3. The free IGF-1 ELISA revealed that 2-DG disrupts the interaction between IGF-1 and IGFBP3 so that the free form of IGF-1 can be released from IGFBP3 binding. However, 2-DG did not appear to interfere with the interaction between IGF-1 and IGF1R, because the combination of 2-DG and IGF-1 also induced a significant elevation in AKT phosphorylation (Fig. 2B, lane 8). Therefore, 2-DG treatment resulted in the release of free IGF-1 to activate IGF1R signaling.
Because IGF-1/IGF1R is also involved in the activation of other signaling pathways (23), we also evaluated their activation status and discovered that 2-DG induced ERK phosphorylation in seven of the NSCLC cell lines and one melanoma cell line. Similar to the findings on activation of PI3K/AKT signaling, 2-DG activated ERK signaling within 15 min at pharmacologically relevant concentrations, and IGF1R was found to be essential for the activation of MEK-ERK signaling by 2-DG.
The activations of AKT and ERK signals have been shown previously to promote cell survival (24, 25). Consistent with these previous studies, we observed an enhancement in growth suppression when 2-DG was combined with either PI3K or MEK inhibitors. However, for each of these, such enhancements were not substantial, suggesting that both signaling pathways contributed to cancer cell survival. We did not evaluate a three-way combination of 2-DG and those inhibitors, LY294002 and U0126, because our phospho-antibody microarray analysis indicated that 2-DG treatment led to a statistically significant elevation of phosphorylation in several other signaling pathways as well, such as JAK/STAT or NFκB, so it was possible that the activation of these other signaling pathways could also contribute to enhanced cell survival. Instead, we focused on the possibility of using a single agent as an inhibitor. Because a chemical inhibitor of IGF1R attenuated the 2-DG-induced phosphorylation in 89% (57 of 64) of the sites having enhanced phosphorylation in our microarray, we believed that a better experimental strategy was to combine 2-DG with IGF1R inhibitor II.
When we used subtoxic doses of 2-DG and IGF1R inhibitor II to treat H460 and H157 cells, this combination inhibited cell proliferation by 90% in both cell lines. In addition, the appearance of caspase-3 cleavage and apoptotic cells indicated that a significant portion of the inhibitory effect on growth was due to apoptosis. Therefore, we believe that we have clearly demonstrated that IGF1R is a key mediator for cell survival after 2-DG treatment, and our in vitro data suggest that the inclusion of an IGF1R inhibitor in 2-DG-based therapies may represent the optimal step to synergistically improve treatment efficacy with this drug.
Supplementary Material
This work was supported, in whole or part, by National Institutes of Health Grants PO1 CA116676-030002 (to W. Z.) and RO1 CA118470-01 (to S.-Y. S.) from NCI.
The on-line version of this article (available at http://www.jbc.org) contains supplemental “Methods,” Figs 1 and 2, and Table 1.
- 2-DG
- 2-deoxyglucose
- EGFR
- epithelial growth factor receptor
- ERK
- extracellular signal-regulated kinase
- IGF
- insulin-like growth factor
- IGF1R
- insulin-like growth factor 1 receptor
- IGFBP
- insulin-like growth factor-binding protein
- PI3K
- phosphatidylinositol 3-kinase
- NSCLC
- non-small cell lung cancer
- 7-ADD
- 7-aminoactinomycin D
- ELISA
- enzyme-linked immunosorbent assay
- siRNA
- small interfering RNA
- JAK/STAT
- Janus kinase/signal transducers and activators of transcription
- MAPK
- mitogen-activated protein kinase
- MEK
- MAPK/ERK kinase.
REFERENCES
- 1.Warburg O. (1956) Science 123,309–314 [DOI] [PubMed] [Google Scholar]
- 2.Fantin V. R., St-Pierre J., Leder P. (2006) Cancer Cell 9,425–434 [DOI] [PubMed] [Google Scholar]
- 3.Christofk H. R., Vander Heiden M. G., Harris M. H., Ramanathan A., Gerszten R. E., Wei R., Fleming M. D., Schreiber S. L., Cantley L. C. (2008) Nature 452,230–233 [DOI] [PubMed] [Google Scholar]
- 4.Pelicano H., Martin D. S., Xu R. H., Huang P. (2006) Oncogene 25,4633–4646 [DOI] [PubMed] [Google Scholar]
- 5.Xu R. H., Pelicano H., Zhou Y., Carew J. S., Feng L., Bhalla K. N., Keating M. J., Huang P. (2005) Cancer Res. 65,613–621 [PubMed] [Google Scholar]
- 6.Maschek G., Savaraj N., Priebe W., Braunschweiger P., Hamilton K., Tidmarsh G. F., De Young L. R., Lampidis T. J. (2004) Cancer Res. 64,31–34 [DOI] [PubMed] [Google Scholar]
- 7.Singh D., Banerji A. K., Dwarakanath B. S., Tripathi R. P., Gupta J. P., Mathew T. L., Ravindranath T., Jain V. (2005) Strahlenther. Onkol. 181,507–514 [DOI] [PubMed] [Google Scholar]
- 8.Mohanti B. K., Rath G. K., Anantha N., Kannan V., Das B. S., Chandramouli B. A., Banerjee A. K., Das S., Jena A., Ravichandran R., Sahi U. P., Kumar R., Kapoor N., Kalia V. K., Dwarakanath B. S., Jain V. (1996) Int. J. Radiat. Oncol. Biol. Phys. 35,103–111 [DOI] [PubMed] [Google Scholar]
- 9.Parniak M., Kalant N. (1985) Can. J. Biochem. Cell Biol. 63,333–340 [DOI] [PubMed] [Google Scholar]
- 10.Karczmar G. S., Arbeit J. M., Toy B. J., Speder A., Weiner M. W. (1992) Cancer Res. 52,71–76 [PubMed] [Google Scholar]
- 11.McComb R. B., Yushok W. D. (1964) Cancer Res. 24,198–205 [PubMed] [Google Scholar]
- 12.Sridhar R., Stroude E. C., Inch W. R. (1979) In Vitro 15,685–690 [DOI] [PubMed] [Google Scholar]
- 13.Kang H. T., Hwang E. S. (2006) Life Sci. 78,1392–1399 [DOI] [PubMed] [Google Scholar]
- 14.Zhong D., Liu X., Schafer-Hales K., Marcus A. I., Khuri F. R., Sun S. Y., Zhou W. (2008) Mol. Cancer Ther. 7,809–817 [DOI] [PubMed] [Google Scholar]
- 15.Zhong D., Liu X., Khuri F. R., Sun S. Y., Vertino P. M., Zhou W. (2008) Cancer Res. 68,7270–7277 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yeatman T. J. (2004) Nat. Rev. Cancer 4,470–480 [DOI] [PubMed] [Google Scholar]
- 17.Pollak M. N., Schernhammer E. S., Hankinson S. E. (2004) Nat. Rev. Cancer 4,505–518 [DOI] [PubMed] [Google Scholar]
- 18.Firth S. M., Baxter R. C. (2002) Endocr. Rev. 23,824–854 [DOI] [PubMed] [Google Scholar]
- 19.Chang Y. S., Wang L., Suh Y. A., Mao L., Karpen S. J., Khuri F. R., Hong W. K., Lee H. Y. (2004) Oncogene 23,6569–6580 [DOI] [PubMed] [Google Scholar]
- 20.Grimberg A. (2003) Cancer Biol. Ther. 2,630–635 [PMC free article] [PubMed] [Google Scholar]
- 21.Ding L., Getz G., Wheeler D. A., Mardis E. R., McLellan M. D., Cibulskis K., Sougnez C., Greulich H., Muzny D. M., Morgan M. B., Fulton L., Fulton R. S., Zhang Q., Wendl M. C., Lawrence M. S., Larson D. E., Chen K., Dooling D. J., Sabo A., Hawes A. C., Shen H., Jhangiani S. N., Lewis L. R., Hall O., Zhu Y., Mathew T., Ren Y., Yao J., Scherer S. E., Clerc K., Metcalf G. A., Ng B., Milosavljevic A., Gonzalez-Garay M. L., Osborne J. R., Meyer R., Shi X., Tang Y., Koboldt D. C., Lin L., Abbott R., Miner T. L., Pohl C., Fewell G., Haipek C., Schmidt H., Dunford-Shore B. H., Kraja A., Crosby S. D., Sawyer C. S., Vickery T., Sander S., Robinson J., Winckler W., Baldwin J., Chirieac L. R., Dutt A., Fennell T., Hanna M., Johnson B. E., Onofrio R. C., Thomas R. K., Tonon G., Weir B. A., Zhao X., Ziaugra L., Zody M. C., Giordano T., Orringer M. B., Roth J. A., Spitz M. R., Wistuba I. I., Ozenberger B., Good P. J., Chang A. C., Beer D. G., Watson M. A., Ladanyi M., Broderick S., Yoshizawa A., Travis W. D., Pao W., Province M. A., Weinstock G. M., Varmus H. E., Gabriel S. B., Lander E. S., Gibbs R. A., Meyerson M., Wilson R. K. (2008) Nature 455,1069–1075 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Deleted in proof
- 23.Casa A. J., Dearth R. K., Litzenburger B. C., Lee A. V., Cui X. (2008) Front. Biosci. 13,3273–3287 [DOI] [PubMed] [Google Scholar]
- 24.Vivanco I., Sawyers C. L. (2002) Nat. Rev. Cancer 2,489–501 [DOI] [PubMed] [Google Scholar]
- 25.Roberts P. J., Der C. J. (2007) Oncogene 26,3291–3310 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.